sft_data.parquet/dataset_info.json

{
  "builder_name": "json",
  "citation": "",
  "config_name": "default",
  "dataset_name": "json",
  "dataset_size": 93267222197,
  "description": "",
  "download_checksums": {
    "/lustre/fs1/portfolios/llmservice/projects/llmservice_fm_text/users/yachen/AceMath/AceReason-SFT/sft_data.jsonl": {
      "num_bytes": 97350183702,
      "checksum": null
    }
  },
  "download_size": 97350183702,
  "features": {
    "category": {
      "dtype": "string",
      "_type": "Value"
    },
    "source": {
      "dtype": "string",
      "_type": "Value"
    },
    "input": {
      "dtype": "string",
      "_type": "Value"
    },
    "output": {
      "dtype": "string",
      "_type": "Value"
    }
  },
  "homepage": "",
  "license": "",
  "size_in_bytes": 190617405899,
  "splits": {
    "train": {
      "name": "train",
      "num_bytes": 93267222197,
      "num_examples": 3970332,
      "shard_lengths": [
        21506,
        21397,
        21454,
        21368,
        21388,
        21469,
        21464,
        21384,
        21643,
        21483,
        21488,
        21391,
        21625,
        21329,
        21422,
        21407,
        21376,
        21241,
        21402,
        21409,
        21614,
        21692,
        21514,
        21311,
        21658,
        21342,
        21438,
        21523,
        21364,
        21286,
        21507,
        21594,
        21224,
        21504,
        21394,
        21522,
        21284,
        21399,
        21389,
        21348,
        21240,
        21369,
        21567,
        21474,
        21538,
        21403,
        21279,
        21340,
        21329,
        21394,
        21321,
        21325,
        21326,
        21383,
        21505,
        21528,
        21562,
        21407,
        21417,
        21344,
        21633,
        21327,
        21438,
        21285,
        21456,
        21472,
        21376,
        21449,
        21368,
        21450,
        21434,
        21412,
        21507,
        21460,
        21426,
        21588,
        21516,
        21378,
        21489,
        21402,
        21432,
        21455,
        21514,
        21504,
        21429,
        21523,
        21311,
        21473,
        21328,
        21411,
        21368,
        21372,
        21523,
        21470,
        21392,
        21381,
        21424,
        21319,
        21484,
        21332,
        21325,
        21512,
        21599,
        21316,
        21477,
        21530,
        21494,
        21482,
        21443,
        21421,
        21423,
        21399,
        21386,
        21580,
        21226,
        21538,
        21434,
        21261,
        21296,
        21448,
        21310,
        21476,
        21350,
        21532,
        21466,
        21540,
        21370,
        21367,
        21381,
        21346,
        21220,
        21357,
        21408,
        21289,
        21519,
        21227,
        21546,
        21353,
        21468,
        21522,
        21327,
        21530,
        21334,
        21414,
        21322,
        21496,
        21404,
        21426,
        21392,
        21384,
        21237,
        21382,
        21236,
        21346,
        21470,
        21472,
        21637,
        21456,
        21473,
        21351,
        21448,
        21418,
        21537,
        21259,
        21447,
        21401,
        21249,
        21413,
        21108,
        21583,
        21538,
        21417,
        21304,
        21515,
        21367,
        21430,
        21262,
        21238,
        21401,
        21408,
        21496,
        21338,
        21357,
        21312,
        21550,
        7895
      ],
      "dataset_name": "json"
    }
  },
  "version": {
    "version_str": "0.0.0",
    "major": 0,
    "minor": 0,
    "patch": 0
  }
}